Download
AgentScope is a production-ready agent framework designed to help developers build, deploy, and scale intelligent agentic applications. It provides essential abstractions that evolve with advancing LLM capabilities, emphasizing reasoning, tool use, and flexible orchestration rather than rigid prompt constraints. With built-in support for ReAct agents, memory, planning, human-in-the-loop control, and real-time voice interaction, developers can create powerful agents in minutes. AgentScope integrates seamlessly with tools, long-term memory systems, MCP, A2A (Agent-to-Agent) protocols, and observability frameworks. It also supports reinforcement learning workflows for tuning agents and improving performance across complex tasks. Deployable locally, serverless in the cloud, or on Kubernetes with OpenTelemetry support, AgentScope is built for both experimentation and production environments.
Features
- Built-in ReAct agents with memory, planning, tool use, human-in-the-loop steering, and structured outputs.
- Real-time voice agents with speech input/output and multi-agent voice interaction support.
- Multi-agent orchestration via MsgHub, pipelines, and dynamic message routing.
- Flexible MCP and A2A protocol integration for tool composition and agent-to-agent communication.
- Agentic reinforcement learning support with sample projects for tuning and evaluation.
- Production-ready deployment options with cloud, Docker, K8s support, and built-in observability (OTel).
Project Samples
